Pulse repetition interval(PRI)modulation recognition and pulse sequence search are significant for effective electronic support measures.In modern electromagnetic environments,different types of inter-pulse slide rada...Pulse repetition interval(PRI)modulation recognition and pulse sequence search are significant for effective electronic support measures.In modern electromagnetic environments,different types of inter-pulse slide radars are highly confusing.There are few available training samples in practical situations,which leads to a low recognition accuracy and poor search effect of the pulse sequence.In this paper,an approach based on bi-directional long short-term memory(BiLSTM)networks and the temporal correlation algorithm for PRI modulation recognition and sequence search under the small sample prerequisite is proposed.The simulation results demonstrate that the proposed algorithm can recognize unilinear,bilinear,sawtooth,and sinusoidal PRI modulation types with 91.43% accuracy and complete the pulse sequence search with 30% missing pulses and 50% spurious pulses under the small sample prerequisite.展开更多
为了提高齿轮故障诊断准确率,解决齿轮故障诊断中数据量大、提取特征困难等问题,构建了齿轮故障诊断系统,采用深度学习方法建立了齿轮故障诊断模型,提出一种基于双层长短时记忆(Binary Long Short Term Memory,Bi LSTM)网络的故障诊断方...为了提高齿轮故障诊断准确率,解决齿轮故障诊断中数据量大、提取特征困难等问题,构建了齿轮故障诊断系统,采用深度学习方法建立了齿轮故障诊断模型,提出一种基于双层长短时记忆(Binary Long Short Term Memory,Bi LSTM)网络的故障诊断方法,并对该方法进行了性能分析和对比实验。结果表明:采用Bi LSTM网络方法进行齿轮故障诊断的准确率达到99.76%,分类效果优于支持向量机、Xg Boost、卷积神经网络和长短时记忆(LSTM)网络等方法,有效地提高了故障诊断精度。展开更多
为实现柔性直流(voltage sourced converter-high voltage direct current,VSC-HVDC)换流阀冷却系统入阀水温的智能预测,文中提出一种基于随机森林(random forest,RF)和双向长短时记忆(bi-directional long short-term memory,BiLSTM)...为实现柔性直流(voltage sourced converter-high voltage direct current,VSC-HVDC)换流阀冷却系统入阀水温的智能预测,文中提出一种基于随机森林(random forest,RF)和双向长短时记忆(bi-directional long short-term memory,BiLSTM)网络混合的柔直换流阀冷却系统入阀水温的预测模型,并以此为基础对柔直换流站阀冷系统的冷却能力进行评估。首先,采用RF算法对由阀冷系统监测变量组成的高维特征集进行重要性分析,筛选出影响入阀水温的重要特征,与历史入阀水温构成输入特征向量。然后,将特征向量输入到BiLSTM预测模型,对模型进行训练并实现对入阀水温的准确预测和冷却能力定量评估。最后,以广东电网某柔直换流站为实例对所提方法进行分析,验证了所提出的基于RF-BiLSTM的混合模型预测精度优于BiLSTM模型、RF模型、支持向量机(support vector machine,SVM)模型和自回归滑动平均模型(auto-regressive and moving average,ARMA)模型,并且实现了冷却能力的定量评估。结果表明该换流站冷却裕量达98%,存在过度冷却、能源浪费的问题,与换流站现场运行情况相符,验证了文中所提方法的有效性和准确性。展开更多
室内区域定位在医疗养老、智慧大楼等领域有着广泛的应用.室内区域定位中最突出的问题是无线电信道效应的动态和不可预测性(如多径传播、信道衰落等)对接收信号强度(received signal strength, RSS)的干扰影响.为了降低无线电的干扰,提...室内区域定位在医疗养老、智慧大楼等领域有着广泛的应用.室内区域定位中最突出的问题是无线电信道效应的动态和不可预测性(如多径传播、信道衰落等)对接收信号强度(received signal strength, RSS)的干扰影响.为了降低无线电的干扰,提出了一种新的基于注意力机制的CNN-BiLSTM的室内区域定位模型,该模型通过捕获粗细粒度特征与定位区域的对应关系来减弱RSS序列对信道变化的依赖.首先,利用卷积神经网络(convolutional neural network, CNN)学习捕捉RSS序列的特征来抽取区域中心点的细粒度特征.然后,利用双向长短时记忆(bidirectional long short-term memory, BiLSTM)网络的存储记忆特性,学习当前与过去RSS序列中隐含区域范围的粗粒度特征.最后,利用注意力机制,通过融合粗细粒度特征,建立RSS序列特征与区域位置的映射关系,获取区域位置信息.真实室内环境下区域定位的实验结果表明,与目前定位效果最好的网格区域综合概率定位模型相比,提出的方法在降低计算复杂度的同时提高了区域定位的准确度和对环境的适应能力.展开更多
基金supported by the National Natural Science Foundation of China(61801143,61971155)the National Natural Science Foundation of Heilongjiang Province(LH2020F019).
文摘Pulse repetition interval(PRI)modulation recognition and pulse sequence search are significant for effective electronic support measures.In modern electromagnetic environments,different types of inter-pulse slide radars are highly confusing.There are few available training samples in practical situations,which leads to a low recognition accuracy and poor search effect of the pulse sequence.In this paper,an approach based on bi-directional long short-term memory(BiLSTM)networks and the temporal correlation algorithm for PRI modulation recognition and sequence search under the small sample prerequisite is proposed.The simulation results demonstrate that the proposed algorithm can recognize unilinear,bilinear,sawtooth,and sinusoidal PRI modulation types with 91.43% accuracy and complete the pulse sequence search with 30% missing pulses and 50% spurious pulses under the small sample prerequisite.
文摘为了提高齿轮故障诊断准确率,解决齿轮故障诊断中数据量大、提取特征困难等问题,构建了齿轮故障诊断系统,采用深度学习方法建立了齿轮故障诊断模型,提出一种基于双层长短时记忆(Binary Long Short Term Memory,Bi LSTM)网络的故障诊断方法,并对该方法进行了性能分析和对比实验。结果表明:采用Bi LSTM网络方法进行齿轮故障诊断的准确率达到99.76%,分类效果优于支持向量机、Xg Boost、卷积神经网络和长短时记忆(LSTM)网络等方法,有效地提高了故障诊断精度。
文摘为实现柔性直流(voltage sourced converter-high voltage direct current,VSC-HVDC)换流阀冷却系统入阀水温的智能预测,文中提出一种基于随机森林(random forest,RF)和双向长短时记忆(bi-directional long short-term memory,BiLSTM)网络混合的柔直换流阀冷却系统入阀水温的预测模型,并以此为基础对柔直换流站阀冷系统的冷却能力进行评估。首先,采用RF算法对由阀冷系统监测变量组成的高维特征集进行重要性分析,筛选出影响入阀水温的重要特征,与历史入阀水温构成输入特征向量。然后,将特征向量输入到BiLSTM预测模型,对模型进行训练并实现对入阀水温的准确预测和冷却能力定量评估。最后,以广东电网某柔直换流站为实例对所提方法进行分析,验证了所提出的基于RF-BiLSTM的混合模型预测精度优于BiLSTM模型、RF模型、支持向量机(support vector machine,SVM)模型和自回归滑动平均模型(auto-regressive and moving average,ARMA)模型,并且实现了冷却能力的定量评估。结果表明该换流站冷却裕量达98%,存在过度冷却、能源浪费的问题,与换流站现场运行情况相符,验证了文中所提方法的有效性和准确性。
